Not Enough B.F. Skinners

What behaviorism taught us about helping troubled kids, and how it lost its way
unhappy child sitting on steps
“We were challenged to observe closely and determine what, in the immediate context, might promote or discourage particular behaviors,” writes W. George Scarlett. Photo: Ingimage
July 7, 2014


“Forty-seven pounds” was the right answer to the question on a quiz that I took in an undergraduate psychology course. The question—“How much did the adolescent girl with anorexia weigh before she was treated with behavioral therapy?”—referred to a famous case in the mid-1960s. A girl with severe anorexia nervosa had starved herself almost to death on a psychiatric ward. In a last-ditch attempt to save her, caregivers moved her from an attractive room to a bare, unattractive one; denied her access to TV, radio and visitors; and informed her that the pleasures she had previously enjoyed would be reinstituted only when she began to eat. It was a classic behavioral intervention: it centered on behaviors rather than emotions, and it provided “reinforcement,” or rewards, for desired behaviors. It worked.

The impact of this case was huge, helping to move behaviorism out of the rat lab and into the real world, where it fostered a new commitment to thinking scientifically. No longer could we be content to discuss what might be going on in a child’s unconscious. Now we were challenged to observe closely and determine what, in the immediate context, might promote or discourage particular behaviors.

And wonderful developments ensued. Ivar Lovaas and his associates at UCLA made history with their behavioral treatments for children with autism. Gerald Patterson’s team at the Oregon Social Training School developed effective programs for adolescents diagnosed with conduct disorder, characterized by behaviors such as lying, stealing, fighting and running away from school.

Good, common-sense advice abounded as well. I remember listening to Harvard’s B.F. Skinner, the behaviorist guru, talk at a conference about how to keep poor urban students from dropping out of high school. He urged his audience to treat this problem as one that stemmed not so much from the students as from their environments—the schools—which weren’t providing appropriate reinforcement.

But soon behavioral approaches started to “break bad,” becoming dehumanizing and over-controlling. By the end of the 1970s, questionable practices were everywhere, especially in schools for troubled children. Critics complained that reward and punishment—sticker charts, praise or food doled out for particular actions and a host of other such techniques—had become the only curriculum, representing a “business deal” approach to educating children.

Unfortunately, behaviorism’s emphasis on scientific thinking seemed to have vanished. Early behaviorists such as Lovaas, Patterson and Skinner didn’t focus on lists of things—like stickers, food or praise—that were supposed to ensure good behavior. Rather, they distinguished among stimuli, which prompt a behavior; reinforcement, which increases a behavior; and punishment, which decreases a behavior. And they knew that what constitutes stimuli, reinforcement or punishment could vary from child to child.

That tailored approach can be invaluable. For example, the staff of a residential program for troubled children, one claiming to follow a behaviorist approach, consulted with me about a child who frequently threw tantrums. They described how he had gone into a fit of rage when told in a reasonable voice that he couldn’t swap part of his lunch for his classmate’s milk because he was lactose intolerant.

When I asked about other cases in which he had thrown tantrums, it became clear that the stimulus for his outburst had been the reasonable voice itself. The one staff member who talked to him in a more sympathetic voice (“Yeah, I know, it’s hard being lactose intolerant”) never set off his tantrums. Normally, being reasonable is a good thing, but not for this boy—as a properly conducted behavioral investigation showed.

Another time I myself made an erroneous assumption, praising a troubled girl who was painting a picture. I had naively thought that my praise would constitute reinforcement, encouraging her constructive activity, but the immediate result was that she tore up her painting and walked away. Later on, after I took the time to learn what really constituted stimuli, reinforcement and punishment for her, I was able to better connect with her.

The point is that behaviorism has much to offer, but only if it includes scientific analysis. Someone once said about psychoanalysis that there are too many Freudians and not enough Freuds. The same could be said about behaviorism: that there are too many Skinnerians and not enough Skinners—practitioners who insist on careful inquiry. It’s worth noting as well that behaviorism need not, indeed must not, replace a commitment to developing positive relationships, designing excellent learning environments and, in general, showing care. No good behaviorist would disagree.

This article first appeared in the Winter 2014 issue of Tufts Magazine.

W. George Scarlett is a senior lecturer in and deputy chair of the Eliot-Pearson Department of Child Study and Human Development.

If You Like This